Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion TRACING.md
Original file line number Diff line number Diff line change
Expand Up @@ -158,6 +158,6 @@ explicitly opening a scope via the `Tracer`.
[otel]: https://opentelemetry.io/
[thread-context]: ./server/src/main/java/org/elasticsearch/common/util/concurrent/ThreadContext.java
[w3c]: https://www.w3.org/TR/trace-context/
[tracing]: ./server/src/main/java/org/elasticsearch/tracing
[tracing]: ./server/src/main/java/org/elasticsearch/telemetry
[agent-config]: https://www.elastic.co/guide/en/apm/agent/java/master/configuration.html
[agent]: https://www.elastic.co/guide/en/apm/agent/java/current/index.html
Original file line number Diff line number Diff line change
Expand Up @@ -50,10 +50,12 @@ class APMJvmOptions {
// by the agent. Don't disable writing to a log file, as the agent will then
// require extra Security Manager permissions when it tries to do something
// else, and it's just painful.
"log_file", "_AGENT_HOME_/../../logs/apm.log",

"log_file", "/Users/przemyslawgomulka/workspace/pgomulka/apm.log",
"log_level", "debug",
// ES does not use auto-instrumentation.
"instrument", "false"
"instrument", "false",
"experimental", "true",
"enable_experimental_instrumentations", "true"
);

/**
Expand Down Expand Up @@ -82,7 +84,7 @@ class APMJvmOptions {
// is doing, leave this value alone.
"log_level", "error",
"application_packages", "org.elasticsearch,org.apache.lucene",
"metrics_interval", "120s",
"metrics_interval", "5s",
"breakdown_metrics", "false",
"central_config", "false"
);
Expand Down Expand Up @@ -316,9 +318,7 @@ static Path findAgentJar(String installDir) throws IOException, UserException {
}

try (var apmStream = Files.list(apmModule)) {
final List<Path> paths = apmStream.filter(
path -> path.getFileName().toString().matches("elastic-apm-agent-\\d+\\.\\d+\\.\\d+\\.jar")
).toList();
final List<Path> paths = apmStream.filter(path -> path.getFileName().toString().matches("elastic-apm-agent-.*.jar")).toList();

if (paths.size() > 1) {
throw new UserException(
Expand Down
5 changes: 5 additions & 0 deletions gradle/verification-metadata.xml
Original file line number Diff line number Diff line change
Expand Up @@ -74,6 +74,11 @@
<sha256 value="3be90268557494f88fe252caf6358d4107aa9349852b8704b1e0ff4831cb74d6" origin="Generated by Gradle"/>
</artifact>
</component>
<component group="co.elastic.apm" name="elastic-apm-agent" version="1.37.0">
<artifact name="elastic-apm-agent-1.37.0.jar">
<sha256 value="10496390d1dcc7c558ae63e115beff9477157e3d219fce686ef5ca59c744dfea" origin="Generated by Gradle"/>
</artifact>
</component>
<component group="co.elastic.logging" name="ecs-logging-core" version="1.2.0">
<artifact name="ecs-logging-core-1.2.0.jar">
<sha256 value="0d6318af29848ea588b1c244834e3e762315c68de2bcfc9f1adf3f0633d22d37" origin="Generated by Gradle"/>
Expand Down
10 changes: 9 additions & 1 deletion modules/apm/build.gradle
Original file line number Diff line number Diff line change
Expand Up @@ -14,11 +14,19 @@ esplugin {

def otelVersion = '1.17.0'

repositories {
maven {
name "sonatype-nexus-snapshots"
url "https://oss.sonatype.org/content/repositories/snapshots"
}
}
dependencies {
implementation "io.opentelemetry:opentelemetry-api:${otelVersion}"
implementation "io.opentelemetry:opentelemetry-context:${otelVersion}"
implementation "io.opentelemetry:opentelemetry-semconv:${otelVersion}-alpha"
runtimeOnly "co.elastic.apm:elastic-apm-agent:1.36.0"
implementation "org.apache.logging.log4j:log4j-api:${versions.log4j}"

runtimeOnly "co.elastic.apm:elastic-apm-agent:1.42.1-SNAPSHOT"
}

tasks.named("dependencyLicenses").configure {
Expand Down
1 change: 1 addition & 0 deletions modules/apm/src/main/java/module-info.java
Original file line number Diff line number Diff line change
Expand Up @@ -16,4 +16,5 @@
requires io.opentelemetry.api;

exports org.elasticsearch.telemetry.apm;
exports org.elasticsearch.telemetry.apm.internal.metrics;
}
Original file line number Diff line number Diff line change
Expand Up @@ -27,6 +27,7 @@
import org.elasticsearch.telemetry.TelemetryProvider;
import org.elasticsearch.telemetry.apm.internal.APMAgentSettings;
import org.elasticsearch.telemetry.apm.internal.APMTelemetryProvider;
import org.elasticsearch.telemetry.apm.internal.metrics.APMMetric;
import org.elasticsearch.telemetry.apm.internal.tracing.APMTracer;
import org.elasticsearch.threadpool.ThreadPool;
import org.elasticsearch.watcher.ResourceWatcherService;
Expand Down Expand Up @@ -97,7 +98,9 @@ public Collection<Object> createComponents(
apmAgentSettings.syncAgentSystemProperties(settings);
apmAgentSettings.addClusterSettingsListeners(clusterService, telemetryProvider.get());

return List.of(apmTracer);
final APMMetric apmMetric = telemetryProvider.get().getMetric();

return List.of(apmTracer, apmMetric);
}

@Override
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,7 @@
import org.elasticsearch.common.settings.Setting;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.core.SuppressForbidden;
import org.elasticsearch.telemetry.apm.internal.metrics.APMMetric;
import org.elasticsearch.telemetry.apm.internal.tracing.APMTracer;

import java.security.AccessController;
Expand All @@ -40,14 +41,21 @@ public class APMAgentSettings {
* Sensible defaults that Elasticsearch configures. This cannot be done via the APM agent
* config file, as then their values could not be overridden dynamically via system properties.
*/
static Map<String, String> APM_AGENT_DEFAULT_SETTINGS = Map.of("transaction_sample_rate", "0.2");
static Map<String, String> APM_AGENT_DEFAULT_SETTINGS = Map.of(
"transaction_sample_rate",
"0.2",
"enable_experimental_instrumentations",
"true"
);

public void addClusterSettingsListeners(ClusterService clusterService, APMTelemetryProvider apmTelemetryProvider) {
final ClusterSettings clusterSettings = clusterService.getClusterSettings();
final APMTracer apmTracer = apmTelemetryProvider.getTracer();
final APMMetric apmMetric = apmTelemetryProvider.getMetric();

clusterSettings.addSettingsUpdateConsumer(APM_ENABLED_SETTING, enabled -> {
apmTracer.setEnabled(enabled);
apmMetric.setEnabled(enabled);
// The agent records data other than spans, e.g. JVM metrics, so we toggle this setting in order to
// minimise its impact to a running Elasticsearch.
this.setAgentSetting("recording", Boolean.toString(enabled));
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -10,19 +10,27 @@

import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.telemetry.TelemetryProvider;
import org.elasticsearch.telemetry.apm.internal.metrics.APMMetric;
import org.elasticsearch.telemetry.apm.internal.tracing.APMTracer;

public class APMTelemetryProvider implements TelemetryProvider {
private final Settings settings;
private final APMTracer apmTracer;
private final APMMetric apmMetric;

public APMTelemetryProvider(Settings settings) {
this.settings = settings;
apmTracer = new APMTracer(settings);
apmMetric = new APMMetric(settings);
}

@Override
public APMTracer getTracer() {
return apmTracer;
}

@Override
public APMMetric getMetric() {
return apmMetric;
}
}
Original file line number Diff line number Diff line change
@@ -0,0 +1,197 @@
/*
* Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
* or more contributor license agreements. Licensed under the Elastic License
* 2.0 and the Server Side Public License, v 1; you may not use this file except
* in compliance with, at your election, the Elastic License 2.0 or the Server
* Side Public License, v 1.
*/

package org.elasticsearch.telemetry.apm.internal.metrics;

import io.opentelemetry.api.GlobalOpenTelemetry;
import io.opentelemetry.api.OpenTelemetry;
import io.opentelemetry.api.metrics.Meter;

import org.elasticsearch.common.component.AbstractLifecycleComponent;
import org.elasticsearch.common.settings.Settings;
import org.elasticsearch.common.util.LazyInitializable;
import org.elasticsearch.telemetry.MetricName;
import org.elasticsearch.telemetry.metric.DoubleCounter;
import org.elasticsearch.telemetry.metric.DoubleGauge;
import org.elasticsearch.telemetry.metric.DoubleHistogram;
import org.elasticsearch.telemetry.metric.DoubleUpDownCounter;
import org.elasticsearch.telemetry.metric.LongCounter;
import org.elasticsearch.telemetry.metric.LongGauge;
import org.elasticsearch.telemetry.metric.LongHistogram;
import org.elasticsearch.telemetry.metric.LongUpDownCounter;

import java.security.AccessController;
import java.security.PrivilegedAction;
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.atomic.AtomicReference;

import static org.elasticsearch.telemetry.apm.internal.APMAgentSettings.APM_ENABLED_SETTING;

public class APMMetric extends AbstractLifecycleComponent implements org.elasticsearch.telemetry.metric.Metric {

private final InstrumentRegistrar<DoubleCounter> doubleCounters = new InstrumentRegistrar<>();
private final InstrumentRegistrar<DoubleUpDownCounter> doubleUpDownCounters = new InstrumentRegistrar<>();
private final InstrumentRegistrar<DoubleGauge> doubleGauges = new InstrumentRegistrar<>();
private final InstrumentRegistrar<DoubleHistogram> doubleHistograms = new InstrumentRegistrar<>();
private final InstrumentRegistrar<LongCounter> longCounters = new InstrumentRegistrar<>();
private final InstrumentRegistrar<LongUpDownCounter> longUpDownCounters = new InstrumentRegistrar<>();
private final InstrumentRegistrar<LongGauge> longGauges = new InstrumentRegistrar<>();
private final InstrumentRegistrar<LongHistogram> longHistograms = new InstrumentRegistrar<>();
private volatile boolean enabled;
private AtomicReference<APMServices> services = new AtomicReference<>();

private final List<SwitchableInstrument> allRegisteredInstruments = new ArrayList<>();

record APMServices(Meter meter, OpenTelemetry openTelemetry) {}

// TODO remove duplication between APMTracer and APMMetric. enabled, create apm services etc
public APMMetric(Settings settings) {
this.enabled = APM_ENABLED_SETTING.get(settings);
}

public void setEnabled(boolean enabled) {
this.enabled = enabled;
if (enabled) {
createApmServices();
allRegisteredInstruments.forEach(si -> si.setEnabled(true));
} else {
allRegisteredInstruments.forEach(si -> si.setEnabled(false));
destroyApmServices();
}

}

@Override
protected void doStart() {
if (enabled) {
createApmServices();
}
}

@Override
protected void doStop() {
destroyApmServices();
}

@Override
protected void doClose() {}

@Override
public <T> DoubleCounter registerDoubleCounter(MetricName name, String description, T unit) {
var counter = OtelDoubleCounter.build(()->services.get().meter, name, description, unit);
doubleCounters.register(counter);
allRegisteredInstruments.add(counter);
return counter;
}

@Override
public <T> DoubleUpDownCounter registerDoubleUpDownCounter(MetricName name, String description, T unit) {
var counter = OtelDoubleUpDownCounter.build(services.get().meter, name, description, unit);
doubleUpDownCounters.register(counter);
allRegisteredInstruments.add(counter);

return counter;
}

@Override
public <T> DoubleGauge registerDoubleGauge(MetricName name, String description, T unit) {
var gauge = OtelDoubleGauge.build(services.get().meter, name, description, unit);
doubleGauges.register(gauge);
allRegisteredInstruments.add(gauge);

return gauge;
}

@Override
public <T> DoubleHistogram registerDoubleHistogram(MetricName name, String description, T unit) {
var histogram = OtelDoubleHistogram.build(services.get().meter, name, description, unit);
doubleHistograms.register(histogram);
allRegisteredInstruments.add(histogram);
return histogram;
}

@Override
public <T> LongCounter registerLongCounter(MetricName name, String description, T unit) {
var lazyCounter = new LazyInitializable<>(
() -> services.get().meter.counterBuilder(name.getRawName()).setDescription(description).setUnit(unit.toString()).build()
);
var counter = OtelLongCounter.build(lazyCounter,services.get().meter, name, description, unit);
longCounters.register(counter);
allRegisteredInstruments.add(counter);

return counter;
}

@Override
public <T> LongUpDownCounter registerLongUpDownCounter(MetricName name, String description, T unit) {

var lazyCounter = new LazyInitializable<>(
() -> services.get().meter.upDownCounterBuilder(name.getRawName()).setDescription(description).setUnit(unit.toString()).build()
);
var counter = OtelLongUpDownCounter.build(lazyCounter,services.get().meter, name, description, unit);
longUpDownCounters.register(counter);
allRegisteredInstruments.add(counter);

return counter;
}

@Override
public <T> LongGauge registerLongGauge(MetricName name, String description, T unit) {

var lazyGauge = new LazyInitializable<>(
() -> services.get().meter.gaugeBuilder(name.getRawName())
.ofLongs()
.setDescription(description)
.setUnit(unit.toString())
.buildObserver()
);

var gauge = OtelLongGauge.build(lazyGauge, services.get().meter, name, description, unit);
longGauges.register(gauge);
allRegisteredInstruments.add(gauge);

return gauge;
}

@Override
public <T> LongHistogram registerLongHistogram(MetricName name, String description, T unit) {

var lazyHistogram = new LazyInitializable<>(
() -> services.get().meter.histogramBuilder(name.getRawName())
.ofLongs()
.setDescription(description)
.setUnit(unit.toString())
.build()
);

var histogram = OtelLongHistogram.build(lazyHistogram,services.get().meter, name, description, unit);
longHistograms.register(histogram);
allRegisteredInstruments.add(histogram);

return histogram;
}

void createApmServices() {
assert this.enabled;
assert this.services.get() == null;

AccessController.doPrivileged((PrivilegedAction<Void>) () -> {
var openTelemetry = GlobalOpenTelemetry.get();
var meter = openTelemetry.getMeter("elasticsearch");

this.services.set(new APMServices(meter, openTelemetry));
return null;
});
}

private void destroyApmServices() {
this.services.set(null);
}

}
Original file line number Diff line number Diff line change
@@ -0,0 +1,35 @@
/*
* Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
* or more contributor license agreements. Licensed under the Elastic License
* 2.0 and the Server Side Public License, v 1; you may not use this file except
* in compliance with, at your election, the Elastic License 2.0 or the Server
* Side Public License, v 1.
*/

package org.elasticsearch.telemetry.apm.internal.metrics;

import org.elasticsearch.common.util.concurrent.ConcurrentCollections;
import org.elasticsearch.telemetry.MetricName;
import org.elasticsearch.telemetry.metric.Instrument;

import java.util.Map;

public class InstrumentRegistrar<T extends Instrument> {
private final Map<MetricName, T> registered = ConcurrentCollections.newConcurrentMap();

void register(T instrument) {
registered.compute(instrument.getName(), (k, v) -> {
if (v != null) {
throw new IllegalStateException(
instrument.getClass().getSimpleName() + "[" + instrument.getName().getRawName() + "] already registered"
);
}

return instrument;
});
}

T get(MetricName name) {
return registered.get(name);
}
}
Loading